Current graphics processing unit (GPU) systems may handle invalidations and flushes by either partially or completely stalling the graphics pipeline, which may be detrimental to performance, based on the frequency of occurrence.